Timing closure & STA

Arrival time

Arrival time is the moment the data signal actually shows up at a point in the circuit, measured from the active clock edge — like the timestamp a delivery truck gets when it finally reaches your door. STA computes it by propagating forward from each startpoint, adding up every gate and wire delay along the way. Because a node can be fed by several converging paths, the tool keeps the latest arrival for setup checks and the earliest for hold checks.

Its partner is the required time — the deadline by which the data must arrive to be safely captured. The whole game of timing analysis reduces to comparing these two numbers: slack equals required time minus arrival time. If arrival beats the deadline, slack is positive and the path is happy; if it arrives late, slack goes negative and you have a violation to fix.

slack = required_time − arrival_time

Arrival times are always given as min/max pairs to bound the real spread of delays caused by process, voltage, and temperature variation — a single nominal number would be a dangerous fiction inside a real silicon corner.

Also called
actual arrival timeAAT